Impacts of US Interest Rates on Growth, Income Distribution, and Macroeconomic Policy Space in Developing Countries: A SFC Supermultiplier Model

ABSTRACT

This paper presents a stock-flow-consistent model in which growth is led by exports, government consumption, and consumption out of interest. It considers domestic and external debt dynamics and gross capital flows. The balance of payments constraint serves as a ceiling but not a floor for growth. Countries may choose to not fully use their external space to decrease public debts or accumulate international reserves. The model is applied to a comparative dynamics exercise to explore how a hike in foreign interest rates may impact growth and income distribution in a developing country under different policy responses. The shock forces the country to apply at least one contractionary macroeconomic policy or lose its reserves. Contractionary monetary policy may be the only tool to avoid external disequilibrium after the shock for countries with open capital accounts. Financial flows may impose trade-offs between short and long run policy objectives. Accumulated international reserves helps maintain expansionary policies and higher average growth rates after external shocks by providing liquidity in foreign currency.

KEYWORDS:

• Supermultiplier
• Stock-flow-consistent modelling
• open economy growth
• macroeconomic policy space
• global financial cycle

JEL CODES:

• E62
• E63
• F32
• F43
• O40

Acknowledgments

This paper was first written as a master’s Thesis supervised by professors Eckhard Hein, Marc Lavoie, and Bruno Tinel. I am grateful to them for instigating discussions and thoughtful advice. It also draws on fruitful discussions with Ricardo Summa, Guilherme Morlin, and Matias Torchinsky. I also thank two anonymous reviewers and the colleagues at the 4th Workshop on Demand-led Growth in IE-UFRJ and at the 16th EAEPE Summer School in Roma Tre for their comments. All remaining errors are my own, of course.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1 According to the IMF, total foreign exchange reserves held by emerging and developing countries went from 714 billion US dollars in 2000 to 7.9 trillion US dollars in 2013.

2 In the 2000s, most developing countries assumed some degree of managed floating regime with interest rate arbitrage, reserve intervention, or capital controls (Ilzetzki, Reinhart, and Rogoff (2022)). Exchange rates are taken exogenously as a simplification to focus on volume changes. Section Four considers the case of fully flexible rates.

3 For matters of simplicity, government consumption and investment were not distinguished. See Deleidi and Mazzucato (2019) for a supermultiplier model which distinguishes public investment.

4 See Dvoskin and Torchinsky Landau (2023) for the consideration of private external debt in a similar model.

5 For simplicity the proportion of installed capital to potential output is assumed to remain constant and equal to unit.

6 All income is taxed proportionally. A progressive taxation could imply higher aggregate demand levels, while the opposite applies for regressive taxation. This is not considered for simplicity.

7 Here the paradox of profits is taken with the profit-share in productive income and not in total income, following Hein (2018).

8 That ‘the expected long-run growth is equal to the known rate of growth of the exogenous component [the autonomous demand]’ (Dutt 2019, p.17). This assumes Harrodian instability out. Nah and Lavoie (2017) propose an adjustment in two different periods with Harrodian instability, this is not done here for matters of simplicity.

9 This is in line with the BPCG literature and its empirical findings that relative price changes explain little to none of the long-run movement of trade (Kaldor 1978; Thirlwall 2012).

10 With Dutt’s (2019) argument, .$\dot{\gamma }=0$. is reached when $g_Z={\gamma }^{\ast \ast }$.

11 This generally points to the returns on the safest assets, Fed bonds, affecting the rentability of assets competiting internationally and captures the usual consideration that interest rates on a country’s external debt are not under its control (Cline and Vernengo 2016). One limitation of the definition, however, is that $i_x$ is assumed completely independent from i, which would not be the case if part of the external debt is held on domestic bonds bought by foreigners. A consideration left for further research.

12 Implicitly this also means that the Central Bank intervenes to sustain exchange rates by reserve intervention, such that all changes in BoP are taken in volumes and not in prices. This is partially relaxed in Section ‘Case 1: Exchange Rate Policy’.

13 Different forms of capital flows are not distinguished, but composition should matter. If the financial account holds more volatile flows, the country will be more sensitive to external shocks. This is captured by $\delta$.

14 Dvoskin and Torchinsky Landau (2023) use the dynamics of the net external debt to international reserves as a proxy for the risk assessment.

15 In the model the ratio represents the BoP result to exports.

16 Thirlwall’s Law can be written as $g_b={\displaystyle \frac{g_x}{\eta }}$. Where $\eta$ is the income elasticity of imports, which is equal to unit in the model the propensity to import is assumed constant. From Equation (15), $g^{\ast \ast }=g_z=g_x=g_b$.

17 ‘Policy constrained’ can also be interpreted as ‘politically constrained’ as in Kalecki (1943 [1984]).

18 Developing countries are often primary good exporters, so their exports have low exchange rate elasticity (Bernat 2015) and due the absence of capital and intermediary goods industries domestically makes the decrease in imports unlikely (Dvoskin and Torchinsky Landau 2023).

19 Note that Equation (26) can be written as $r^{\ast \ast }={\displaystyle \frac{X-M+(g_x-i_x)d^{\ast \ast }}{g_x-i_f}}$, which highlights that when $i_x>g_x>i_f$, ${\displaystyle \frac{\mathrm{\partial }{r}^{\ast \ast }}{\mathrm{\partial }{d}^{\ast \ast }}={\displaystyle \frac{g_x-i_x}{g_x-i_f}<0}}$. For stability $g_x>i_f$ was assumed. Since $i_x>i_f$, $i_x>g_x>i_f$ is possible.

20 ${\displaystyle \frac{\mathrm{\partial }{\left({\displaystyle \frac{Y_P}{X}}\right)}^{\ast }}{\mathrm{\partial }e}={\displaystyle \frac{\mathrm{\partial }\pi }{\mathrm{\partial }e}\ast {\displaystyle \frac{\mathrm{\partial }{\left({\displaystyle \frac{Y_P}{X}}\right)}^{\ast }}{\mathrm{\partial }\pi }={\displaystyle \frac{\mathrm{\partial }\pi }{\mathrm{\partial }e}\ast {\displaystyle \frac{\mathrm{\partial }\left({\displaystyle \frac{\gamma -{\gamma }_u{u}_n+z}{(1-\alpha )z(s_r\pi +{\varphi }_u+\tau -{\gamma }_u)}}\right)}{\mathrm{\partial }\pi }={\displaystyle \frac{\mathrm{\partial }\pi }{\mathrm{\partial }e}\ast {\displaystyle \frac{-s_r(\gamma -{\gamma }_u{u}_n+z)}{(1-\alpha )z{(s_r\pi +{\varphi }_u+\tau -{\gamma }_u)}^2}<0}}}}}}}$, since it was defined that ${\displaystyle \frac{\mathrm{\partial }\pi }{\mathrm{\partial }e}>0}$. In the long-run, the effect on z** must be considered.${\displaystyle \frac{\mathrm{\partial }{\left({\displaystyle \frac{Y_P}{X}}\right)}^{\ast \ast }}{\mathrm{\partial }e}={\displaystyle \frac{\mathrm{\partial }\pi }{\mathrm{\partial }e}\ast \left[{\displaystyle \frac{-s_r({\gamma }^{\ast \ast }-{\gamma }_u{u}_n+z^{\ast \ast })}{(1-\alpha )z^{\ast \ast }{(s_r\pi +{\varphi }_u+\tau -{\gamma }_u)}^2}+{\displaystyle \frac{\mathrm{\partial }{z}^{\ast \ast }}{\mathrm{\partial }e}}}\right]=}}$ ${\displaystyle \frac{\mathrm{\partial }\pi }{\mathrm{\partial }e}\ast \left[{\displaystyle \frac{-s_r(\gamma -{\gamma }_u{u}_n+z)}{(1-\alpha )z{(s_r\pi +{\varphi }_u+\tau -{\gamma }_u)}^2}-{\displaystyle \frac{s_r{u}_n({\gamma }^{\ast \ast }-{\gamma }_u{u}_n)}{(1-\alpha )z^{\ast \ast 2}(s_r\pi +{\varphi }_u+\tau -{\gamma }_u)}}}\right]<0}$.

21 Fromm Equations (7) and (9), the condition is that the autonomous demand excluding consumption out of interest is positive, $\gamma +x+\vartheta -{\gamma }_u{u}_n>0$. A condition that should be easily met by all economies.

22 The possibility of change in the tax rate is not considered, assuming changing tax rate policy takes longer time. Generally, the tax rate affects BoP negatively in the medium-run, while long-run results would depend on the parameters.

23 ${\displaystyle \frac{\mathrm{\partial }{\left({\displaystyle \frac{Y_P}{X}}\right)}^{\ast }}{\mathrm{\partial }\alpha }={\displaystyle \frac{\mathrm{\partial }{\left({\displaystyle \frac{Y_P}{X}}\right)}^{\ast \ast }}{\mathrm{\partial }\alpha }={\displaystyle \frac{\mathrm{\partial }\left({\displaystyle \frac{({\gamma }^{\ast \ast }-{\gamma }_u{u}_n+z^{\ast \ast })}{(1-\alpha )z^{\ast \ast }(s_r\pi +{\varphi }_u+\tau -{\gamma }_u)}}\right)}{\mathrm{\partial }\alpha }={\displaystyle \frac{({\gamma }^{\ast \ast }-{\gamma }_u{u}_n+z^{\ast \ast })}{{(1-\alpha )}^2{z}^{\ast \ast }(s_r\pi +{\varphi }_u+\tau -{\gamma }_u)}>0}}}}$

24 See Botta et al. (2023) for more on this.